Mercurial > libavcodec.hg
diff i386/mpegvideo_mmx_template.c @ 2979:bfabfdf9ce55 libavcodec
COSMETICS: tabs --> spaces, some prettyprinting
| author | diego |
|---|---|
| date | Thu, 22 Dec 2005 01:10:11 +0000 |
| parents | ef2149182f1c |
| children | 0b546eab515d |
line wrap: on
line diff
--- a/i386/mpegvideo_mmx_template.c Wed Dec 21 17:50:40 2005 +0000 +++ b/i386/mpegvideo_mmx_template.c Thu Dec 22 01:10:11 2005 +0000 @@ -21,26 +21,26 @@ #undef PMAXW #ifdef HAVE_MMX2 #define SPREADW(a) "pshufw $0, " #a ", " #a " \n\t" -#define PMAXW(a,b) "pmaxsw " #a ", " #b " \n\t" +#define PMAXW(a,b) "pmaxsw " #a ", " #b " \n\t" #define PMAX(a,b) \ - "pshufw $0x0E," #a ", " #b " \n\t"\ - PMAXW(b, a)\ - "pshufw $0x01," #a ", " #b " \n\t"\ - PMAXW(b, a) + "pshufw $0x0E," #a ", " #b " \n\t"\ + PMAXW(b, a)\ + "pshufw $0x01," #a ", " #b " \n\t"\ + PMAXW(b, a) #else #define SPREADW(a) \ - "punpcklwd " #a ", " #a " \n\t"\ - "punpcklwd " #a ", " #a " \n\t" + "punpcklwd " #a ", " #a " \n\t"\ + "punpcklwd " #a ", " #a " \n\t" #define PMAXW(a,b) \ - "psubusw " #a ", " #b " \n\t"\ - "paddw " #a ", " #b " \n\t" + "psubusw " #a ", " #b " \n\t"\ + "paddw " #a ", " #b " \n\t" #define PMAX(a,b) \ - "movq " #a ", " #b " \n\t"\ - "psrlq $32, " #a " \n\t"\ - PMAXW(b, a)\ - "movq " #a ", " #b " \n\t"\ - "psrlq $16, " #a " \n\t"\ - PMAXW(b, a) + "movq " #a ", " #b " \n\t"\ + "psrlq $32, " #a " \n\t"\ + PMAXW(b, a)\ + "movq " #a ", " #b " \n\t"\ + "psrlq $16, " #a " \n\t"\ + PMAXW(b, a) #endif @@ -71,18 +71,18 @@ if (!s->h263_aic) { #if 1 asm volatile ( - "mul %%ecx \n\t" - : "=d" (level), "=a"(dummy) - : "a" ((block[0]>>2) + q), "c" (inverse[q<<1]) + "mul %%ecx \n\t" + : "=d" (level), "=a"(dummy) + : "a" ((block[0]>>2) + q), "c" (inverse[q<<1]) ); #else asm volatile ( - "xorl %%edx, %%edx \n\t" - "divw %%cx \n\t" - "movzwl %%ax, %%eax \n\t" - : "=a" (level) - : "a" ((block[0]>>2) + q), "c" (q<<1) - : "%edx" + "xorl %%edx, %%edx \n\t" + "divw %%cx \n\t" + "movzwl %%ax, %%eax \n\t" + : "=a" (level) + : "a" ((block[0]>>2) + q), "c" (q<<1) + : "%edx" ); #endif } else @@ -103,94 +103,94 @@ if((s->out_format == FMT_H263 || s->out_format == FMT_H261) && s->mpeg_quant==0){ asm volatile( - "movd %%"REG_a", %%mm3 \n\t" // last_non_zero_p1 + "movd %%"REG_a", %%mm3 \n\t" // last_non_zero_p1 SPREADW(%%mm3) - "pxor %%mm7, %%mm7 \n\t" // 0 - "pxor %%mm4, %%mm4 \n\t" // 0 - "movq (%2), %%mm5 \n\t" // qmat[0] - "pxor %%mm6, %%mm6 \n\t" - "psubw (%3), %%mm6 \n\t" // -bias[0] - "mov $-128, %%"REG_a" \n\t" - ".balign 16 \n\t" - "1: \n\t" - "pxor %%mm1, %%mm1 \n\t" // 0 - "movq (%1, %%"REG_a"), %%mm0 \n\t" // block[i] - "pcmpgtw %%mm0, %%mm1 \n\t" // block[i] <= 0 ? 0xFF : 0x00 - "pxor %%mm1, %%mm0 \n\t" - "psubw %%mm1, %%mm0 \n\t" // ABS(block[i]) - "psubusw %%mm6, %%mm0 \n\t" // ABS(block[i]) + bias[0] - "pmulhw %%mm5, %%mm0 \n\t" // (ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16 - "por %%mm0, %%mm4 \n\t" - "pxor %%mm1, %%mm0 \n\t" - "psubw %%mm1, %%mm0 \n\t" // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i]) - "movq %%mm0, (%5, %%"REG_a") \n\t" - "pcmpeqw %%mm7, %%mm0 \n\t" // out==0 ? 0xFF : 0x00 - "movq (%4, %%"REG_a"), %%mm1 \n\t" - "movq %%mm7, (%1, %%"REG_a") \n\t" // 0 - "pandn %%mm1, %%mm0 \n\t" - PMAXW(%%mm0, %%mm3) - "add $8, %%"REG_a" \n\t" - " js 1b \n\t" - PMAX(%%mm3, %%mm0) - "movd %%mm3, %%"REG_a" \n\t" - "movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1 - : "+a" (last_non_zero_p1) + "pxor %%mm7, %%mm7 \n\t" // 0 + "pxor %%mm4, %%mm4 \n\t" // 0 + "movq (%2), %%mm5 \n\t" // qmat[0] + "pxor %%mm6, %%mm6 \n\t" + "psubw (%3), %%mm6 \n\t" // -bias[0] + "mov $-128, %%"REG_a" \n\t" + ".balign 16 \n\t" + "1: \n\t" + "pxor %%mm1, %%mm1 \n\t" // 0 + "movq (%1, %%"REG_a"), %%mm0 \n\t" // block[i] + "pcmpgtw %%mm0, %%mm1 \n\t" // block[i] <= 0 ? 0xFF : 0x00 + "pxor %%mm1, %%mm0 \n\t" + "psubw %%mm1, %%mm0 \n\t" // ABS(block[i]) + "psubusw %%mm6, %%mm0 \n\t" // ABS(block[i]) + bias[0] + "pmulhw %%mm5, %%mm0 \n\t" // (ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16 + "por %%mm0, %%mm4 \n\t" + "pxor %%mm1, %%mm0 \n\t" + "psubw %%mm1, %%mm0 \n\t" // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i]) + "movq %%mm0, (%5, %%"REG_a") \n\t" + "pcmpeqw %%mm7, %%mm0 \n\t" // out==0 ? 0xFF : 0x00 + "movq (%4, %%"REG_a"), %%mm1 \n\t" + "movq %%mm7, (%1, %%"REG_a") \n\t" // 0 + "pandn %%mm1, %%mm0 \n\t" + PMAXW(%%mm0, %%mm3) + "add $8, %%"REG_a" \n\t" + " js 1b \n\t" + PMAX(%%mm3, %%mm0) + "movd %%mm3, %%"REG_a" \n\t" + "movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1 + : "+a" (last_non_zero_p1) : "r" (block+64), "r" (qmat), "r" (bias), "r" (inv_zigzag_direct16+64), "r" (temp_block+64) ); // note the asm is split cuz gcc doesnt like that many operands ... asm volatile( - "movd %1, %%mm1 \n\t" // max_qcoeff - SPREADW(%%mm1) - "psubusw %%mm1, %%mm4 \n\t" - "packuswb %%mm4, %%mm4 \n\t" - "movd %%mm4, %0 \n\t" // *overflow + "movd %1, %%mm1 \n\t" // max_qcoeff + SPREADW(%%mm1) + "psubusw %%mm1, %%mm4 \n\t" + "packuswb %%mm4, %%mm4 \n\t" + "movd %%mm4, %0 \n\t" // *overflow : "=g" (*overflow) : "g" (s->max_qcoeff) ); }else{ // FMT_H263 asm volatile( - "movd %%"REG_a", %%mm3 \n\t" // last_non_zero_p1 + "movd %%"REG_a", %%mm3 \n\t" // last_non_zero_p1 SPREADW(%%mm3) - "pxor %%mm7, %%mm7 \n\t" // 0 - "pxor %%mm4, %%mm4 \n\t" // 0 - "mov $-128, %%"REG_a" \n\t" - ".balign 16 \n\t" - "1: \n\t" - "pxor %%mm1, %%mm1 \n\t" // 0 - "movq (%1, %%"REG_a"), %%mm0 \n\t" // block[i] - "pcmpgtw %%mm0, %%mm1 \n\t" // block[i] <= 0 ? 0xFF : 0x00 - "pxor %%mm1, %%mm0 \n\t" - "psubw %%mm1, %%mm0 \n\t" // ABS(block[i]) - "movq (%3, %%"REG_a"), %%mm6 \n\t" // bias[0] - "paddusw %%mm6, %%mm0 \n\t" // ABS(block[i]) + bias[0] - "movq (%2, %%"REG_a"), %%mm5 \n\t" // qmat[i] - "pmulhw %%mm5, %%mm0 \n\t" // (ABS(block[i])*qmat[0] + bias[0]*qmat[0])>>16 - "por %%mm0, %%mm4 \n\t" - "pxor %%mm1, %%mm0 \n\t" - "psubw %%mm1, %%mm0 \n\t" // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i]) - "movq %%mm0, (%5, %%"REG_a") \n\t" - "pcmpeqw %%mm7, %%mm0 \n\t" // out==0 ? 0xFF : 0x00 - "movq (%4, %%"REG_a"), %%mm1 \n\t" - "movq %%mm7, (%1, %%"REG_a") \n\t" // 0 - "pandn %%mm1, %%mm0 \n\t" - PMAXW(%%mm0, %%mm3) - "add $8, %%"REG_a" \n\t" - " js 1b \n\t" - PMAX(%%mm3, %%mm0) - "movd %%mm3, %%"REG_a" \n\t" - "movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1 - : "+a" (last_non_zero_p1) + "pxor %%mm7, %%mm7 \n\t" // 0 + "pxor %%mm4, %%mm4 \n\t" // 0 + "mov $-128, %%"REG_a" \n\t" + ".balign 16 \n\t" + "1: \n\t" + "pxor %%mm1, %%mm1 \n\t" // 0 + "movq (%1, %%"REG_a"), %%mm0 \n\t" // block[i] + "pcmpgtw %%mm0, %%mm1 \n\t" // block[i] <= 0 ? 0xFF : 0x00 + "pxor %%mm1, %%mm0 \n\t" + "psubw %%mm1, %%mm0 \n\t" // ABS(block[i]) + "movq (%3, %%"REG_a"), %%mm6 \n\t" // bias[0] + "paddusw %%mm6, %%mm0 \n\t" // ABS(block[i]) + bias[0] + "movq (%2, %%"REG_a"), %%mm5 \n\t" // qmat[i] + "pmulhw %%mm5, %%mm0 \n\t" // (ABS(block[i])*qmat[0] + bias[0]*qmat[0])>>16 + "por %%mm0, %%mm4 \n\t" + "pxor %%mm1, %%mm0 \n\t" + "psubw %%mm1, %%mm0 \n\t" // out=((ABS(block[i])*qmat[0] - bias[0]*qmat[0])>>16)*sign(block[i]) + "movq %%mm0, (%5, %%"REG_a") \n\t" + "pcmpeqw %%mm7, %%mm0 \n\t" // out==0 ? 0xFF : 0x00 + "movq (%4, %%"REG_a"), %%mm1 \n\t" + "movq %%mm7, (%1, %%"REG_a") \n\t" // 0 + "pandn %%mm1, %%mm0 \n\t" + PMAXW(%%mm0, %%mm3) + "add $8, %%"REG_a" \n\t" + " js 1b \n\t" + PMAX(%%mm3, %%mm0) + "movd %%mm3, %%"REG_a" \n\t" + "movzb %%al, %%"REG_a" \n\t" // last_non_zero_p1 + : "+a" (last_non_zero_p1) : "r" (block+64), "r" (qmat+64), "r" (bias+64), "r" (inv_zigzag_direct16+64), "r" (temp_block+64) ); // note the asm is split cuz gcc doesnt like that many operands ... asm volatile( - "movd %1, %%mm1 \n\t" // max_qcoeff - SPREADW(%%mm1) - "psubusw %%mm1, %%mm4 \n\t" - "packuswb %%mm4, %%mm4 \n\t" - "movd %%mm4, %0 \n\t" // *overflow + "movd %1, %%mm1 \n\t" // max_qcoeff + SPREADW(%%mm1) + "psubusw %%mm1, %%mm4 \n\t" + "packuswb %%mm4, %%mm4 \n\t" + "movd %%mm4, %0 \n\t" // *overflow : "=g" (*overflow) : "g" (s->max_qcoeff) );